Download Text S1.

 Enke, Dong, and Bender Supporting Information
Supporting Information
The dcl3 and drm1 drm2 mutations impair acquisition of new PAI2 5meC
To test the effects of dcl mutations in establishing new PAI 5meC imprints triggered by
the PAI1-PAI4 transcribed IR we used a previously developed assay (Figure S2). In this assay, a
pai1-PAI4 missense mutant variant of the 5meC trigger locus is combined by crossing with a
naïve unmethylated target PAI2 gene from a strain that lacks a PAI1-PAI4 IR such as Columbia
(Col) or Landsberg erecta (Ler) [1,2]. Because the PAI3 and PAI4 genes contain polymorphisms
that inactivate enzyme function [3], in the resulting hybrid strain the target PAI2 gene is the only
source of fully functional PAI enzyme. Therefore PAI2 5meC-mediated transcriptional silencing
can be detected by acquisition of a PAI-deficient blue fluorescent phenotype caused by
accumulation of the tryptophan precursor anthranilate. In a wild type background, the pai1-PAI4
transcribed IR triggers accumulation of 5meC and reduced expression of the target PAI2 gene
sufficient for blue fluorescence by the F2 or F3 (S1) generation. The 5meC density on PAI2
triggered by the PAI IR locus increases progressively over subsequent generations of inbreeding
by self-pollination [4].
We performed a similar assay combining pai1-PAI4 and unmethylated PAI2 in hybrids
where both parents carried single dcl2, dcl3, or dcl4 mutations. Independent hybrids generated in
the dcl2 or dcl4 mutant backgrounds (two lines for each genotype) acquired blue fluorescence by
the F2 generation, similarly to hybrids generated in a wild type background. The blue
fluorescence in the dcl2 and dcl4 hybrids corresponded to acquisition of 5meC on the target PAI2
gene, as assessed by HincII DNA gel blot assays on DNA from F5 (S3) generation plants (Figure
S2). In contrast, none of six independent hybrid lines generated in the dcl3 mutant background
displayed blue fluorescence in the F2 generation, and none of these lines segregated blue
Enke, Dong, and Bender Supporting Information
fluorescent progeny until the F6 (S4) generation. HincII DNA gel blot analysis of dcl3 hybrids
indicated that non-fluorescent F5 (S3) plants carried a low level of PAI2 non-CG methylation
whereas fluorescent F8 (S6) plants carried PAI2 non-CG methylation at a similar density to
fluorescent wild type, dcl2, or dcl4 hybrids. These results indicate that DCL3 is the major dicer
involved in initiation of PAI 5meC, but that other dicers can contribute at a less efficient level in
the absence of DCL3.
We also used the hybrid plant assay to test the effects of the drm1 drm2 mutations on
initiation of PAI2 5meC imprints. None of six independent drm1 drm2 hybrid lines segregated
blue fluorescent individuals or displayed detectable PAI2 non-CG methylation as assessed by
HincII DNA gel blot even by the F8 (S6) generation (Figure S2). These results indicate that
DRM1/DRM2 are required for initiation of PAI2 5meC, even though they are dispensable for
maintaining existing PAI 5meC (Figure 1). In a previous study we found that a cmt3 mutation
also prevents initiation of PAI2 5meC in the hybrid plant assay [1]. Taken together, our results
support the view that DRM1/DRM2 initiate a low level of PAI 5meC, which is then amplified
and maintained through the SUVH/CMT3 and MET1 pathways.
Enke, Dong, and Bender Supporting Information
Materials and Methods for Supporting Information
Assay for initiation of PAI2 5meC
To analyze the effects of dcl mutations on initiation of PAI2 5meC imprints, the Col allele
of dcl2-1, dcl3-1, or dcl4-2 was crossed with the same allele in the Ws pai1 background. F2
progeny were scored with PCR-based genotype markers to find individuals homozygous for the
Ws pai1-PAI4 locus and homozygous for the Col PAI2 locus (Table S1). For dcl2 and dcl4 two
independent lines with the desired genotype were analyzed for each cross. For dcl3 and drm1
drm2, six independent lines with the desired genotype were analyzed for each cross. In each
subsequent generation of each inbred F2 line, progeny from two independent plants were assessed
for blue fluorescence under UV light. Because the drm1 drm2 alleles were originally isolated in
Ws, we performed a series of crosses to move these alleles into a background where the PAI2
gene had not been even transiently exposed to Ws PAI1-PAI4: 1) Ws drm1 drm2 was crossed
with Col, and F2 plants were scored with PCR-based genotype markers to find drm1 drm2
individuals homozygous for Col PAI1 and Col PAI2, 2) three F2 individuals with the desired
genotype from the first-round cross were each crossed with Ler, and F2 plants were scored with
PCR-based genotype markers to find drm1 drm2 individuals homozygous for Ler PAI2, 3) one F2
individual with the desired genotype from each second-round cross was crossed with a Ws pai1
drm1 drm2 strain, and F2 plants were scored with PCR-based genotype markers to find
individuals homozygous for the Ws pai1-PAI4 locus and homozygous for the Ler PAI2 locus,
and 4) two individuals with the desired genotype from each of the three third-round crosses were
inbred by self-pollination for a total of six independent lines.
Enke, Dong, and Bender Supporting Information
References for Supporting Information
1. Malagnac F, Bartee L, Bender J (2002) An Arabidopsis SET domain protein required for
maintenance but not establishment of DNA methylation. EMBO J 21: 6842-6852.
2. Melquist S, Bender J (2004) An internal rearrangement in an Arabidopsis inverted repeat locus
impairs DNA methylation triggered by the locus. Genetics 166: 437-448.
3. Melquist S, Luff B, Bender J (1999) Arabidopsis PAI gene arrangements, cytosine methylation
and expression. Genetics 153: 401-413.
4. Luff B, Pawlowski L, Bender J (1999) An inverted repeat triggers cytosine methylation of
identical sequences in Arabidopsis. Mol Cell 3: 505-511.
5. Bender J, Fink GR (1995) Epigenetic control of an endogenous gene family is revealed by a
novel blue fluorescent mutant of Arabidopsis. Cell 83: 725-734.